Date: Sept 2018

This file contains instructions to reproduce the results in the paper
"Impact of Violent Crime on Risk Aversion: Evidence from the Mexican Drug War" by Ryan Brown, Veronica Montalva, Duncan Thomas and Andrea Velasquez     

The authors provide the data, program estimation file, and replication tables (which are created by the program estimation file).

The STATA program .do file conf_pref_ReSTAT.do creates Table 1-Table 5 in the paper using 
conf_pref_ReSTAT.dta. These results were originally generated using STATA/SE 14.2. conf_pref_ReSTAT.do requires the commands, outreg2 and tabout, which are available from the Boston College Statistical Software Components Archive (SSC). They are installed at the top of the analysis file.

conf_pref_ReSTAT.dta was generated from the Mexican Family Life Survey, which can be found at http://www.ennvih-mxfls.org/english/index.html. conf_pref_ReSTAT.dta contains variables constructed using the baseline, ENNVIH-1 (2002), first follow-up, ENNVIH-2 (2005-2006), and second follow up, ENNVIH-3 (2009-2012), of the Mexican Family Life Survey.
	
The output of conf_pref_ReSTAT.do are 5 tables to replicate the primary tables in the paper:
	Table_1.out is Table 1
	Table_2.out is Table 2
	Table_3.out is Table 3
	Table_4.out is Table 4
	Table_5.out is Table 5


Data Dictionary (varname "definition"): 

nchild_p "# of children recorded in previous wave, set to mean if missing"
hhsize_p "household size in previous wave, set to mean if missing"
married_p "=1 if married or cohabiting in previous wave, -10 if missing, 0 otherwise"
emp "=1 if worked last week, 0 otherwise"
emp_p "=1 if worked last week in previous wave,-10 if missing, 0 otherwise"
rural_p "=1 if live in a rural locality in previous wave, -10 if missing, 0 otherwise
self_p "=1 if self-employed in previous wave, -10 if missing, 0 otherwise"
yeduc_p "years of education in previous wave, set to mean if missing"
ohhm_kids_p "number of children of other HH members in previous wave,  set to mean if missing"
qrearn_p "quartic root of last months earnings in previous wave,  set to mean if missing"
qrpce_p "quartic root of per capita expenditure in previous wave,  set to mean if missing"
risk2 "risk index"
m1risk2av "=1 if most risk averse, 0 otherwise"
mes_int "month of interview"
ano_int "year of interview"
wave "=5 if observation is from MxFLS2, =9 if observation is from MxFLS3"
hr "homicide rate in MxFLS2 municipality per 10,000 people over the last 12 months"
rural_05 "=1 if live in a rural locality in MxFLS2"
scared_night "=1 if feel scared of being attacked during the night, 0 otherwise"
safe_c5_less "=1 if feel less safe than 5 years ago, 0 otherwise"
scared_night_p "=1 if scared of attack at night in previous wave, -10 if missing, 0 otherwise"
bp_sys_alt "systolic blood pressure, MxFLS1 value used for MxFLS2 obs."
qrpce0509_B25 "=1 if in bottom quartile of change in per capita expenditure, 0 otherwise"
dmale "=1 if male, 0 if female"
quar_smtot2 "quartic root of emotional wellbeing index"
married_p_M "=1 if married_p is missing, 0 otherwise"
emp_p_M "=1 if emp_p is missing, 0 otherwise"
self_p_M "=1 if self_p is missing, 0 otherwise"
scared_night_p_M "=1 if scared_night_p is missing, 0 otherwise"
yeduc_p_M "=1 if yeduc_p is missing, 0 otherwise"
nchild_p_M "=1 if nchild_p is missing, 0 otherwise"
qrearn_p_M "=1 if qrearn_p is missing, 0 otherwise"
qrpce_p_M "=1 if qrpce_p is missing, 0 otherwise"
age_05 "age in MxFLS2"
self_05 "=1 if self-employed in MxFLS2, 0 otherwise"
yeduc_05_B25 "=1 if in bottom quartile of years of education in MxFLS2,0 otherwise"
qrpce_05_B25 "=1 if in bottom quartile of per capita expenditure in MxFLS2, 0 otherwise"
unique_id "unique individual identifier"
mpio05_id "municipality of residence in MxFLS2 identifier"
mpio_id "current municipality of residence identifier"
emp_100 "=100 if employed, 0 otherwise"
qrpce_B25_100 "=100 if in bottom quartile of per capita expenditure, 0 otherwise"
safe_c5_less_100 "=100 if feel less safe than 5 years ago, 0 otherwise"
scared_night_100 "=100 if feel scared of being attacked during the night, 0 otherwise"



	